Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 28
    Prev
    Next
    INDEX
    Explanations

    mentions of medical terms and organizations associated with healthcare

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    medical and healthcare-related terms and organizations, particularly references to Asante healthcare system.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    This neuron detects organization and brand names, particularly those in healthcare, medical, and commercial domains.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    specific names and terms associated with organizations, places, or titles, often appearing in proper nouns.

    oai_token-act-pair · gpt-4oTriggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    hésite
    -0.35
     detuvo
    -0.34
    andangan
    -0.34
    hysema
    -0.33
     tilgjenge
    -0.32
    cuál
    -0.32
     bague
    -0.31
    ợp
    -0.31
    катери
    -0.31
     uſed
    -0.31
    POSITIVE LOGITS
    AnchorStyles
    0.67
     BoxFit
    0.65
    ValueStyle
    0.62
    ronpa
    0.62
    Datuak
    0.55
    stateMutability
    0.54
     CURLOPT
    0.52
     GlobalKey
    0.52
     ब्रेकडाउन
    0.52
     wireType
    0.51
    Activations Density 0.986%

    No Known Activations